CN105487544B - Method is surrounded and seize in multirobot angle control based on fuzzy inference system - Google Patents
Method is surrounded and seize in multirobot angle control based on fuzzy inference system Download PDFInfo
- Publication number
- CN105487544B CN105487544B CN201610032491.7A CN201610032491A CN105487544B CN 105487544 B CN105487544 B CN 105487544B CN 201610032491 A CN201610032491 A CN 201610032491A CN 105487544 B CN105487544 B CN 105487544B
- Authority
- CN
- China
- Prior art keywords
- seize
- surrounding
- person
- runaway
- robot
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 27
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 16
- 238000013459 approach Methods 0.000 claims abstract description 15
- 238000002474 experimental method Methods 0.000 claims abstract description 13
- 230000015572 biosynthetic process Effects 0.000 claims description 10
- 230000008602 contraction Effects 0.000 claims description 7
- 238000013461 design Methods 0.000 claims description 5
- 230000009467 reduction Effects 0.000 claims description 4
- 239000002245 particle Substances 0.000 claims description 3
- 230000007423 decrease Effects 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 8
- 238000004088 simulation Methods 0.000 abstract description 4
- 230000006870 function Effects 0.000 description 13
- 230000008569 process Effects 0.000 description 7
- 238000011160 research Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 238000009826 distribution Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 210000000987 immune system Anatomy 0.000 description 2
- 230000035772 mutation Effects 0.000 description 2
- 241001465754 Metazoa Species 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 210000003719 b-lymphocyte Anatomy 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000000571 coke Substances 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 239000010977 jade Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000002028 premature Effects 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0221—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving a learning process
Abstract
Method is surrounded and seize in multirobot angle control based on fuzzy inference system, and this method realizes that multirobot surrounds and seize strategy using two layers of fuzzy reasoning;First layer fuzzy inference system is that decision-making level selects corresponding multirobot strategy for being identified to surrounding and seize task status;When the person of surrounding and seize is in search condition, the output of decision-making level's fuzzy reasoningSearch, the person's of surrounding and seize execution search strategy;When the person of surrounding and seize is in proximity state, decision-making level's outputApproach, i.e., close to strategy;When the person of surrounding and seize, which is in, surrounds and seize state, decision-making level's outputSurround, the person of surrounding and seize carries out surrounding and seize strategy.It is tested using simulated program, the environment of simulated program proportionally reduces according to the parameter in actual environment, and the movement of robot meets kinematics model constraint.Multi simulation running experiment is carried out under different primary condition, is demonstrated the feasibility of algorithm, is achieved preferable effect.
Description
Technical field:
The multirobot that the present invention relates to a kind of in a dynamic environment based on angle control surrounds and seize strategy, specially a kind of
Method is surrounded and seize in multirobot angle control based on fuzzy inference system.
Background technology:
Due to multi-robot system have distributivity spatially, diversity functionally, execute task when concurrency,
The incomparable superiority of stronger fault-tolerant ability and lower economic cost Deng Dan robots so that multi-robot system is in recent years
To become the hot spot of people's extensive concern.The main problem of multi-robot system research includes group structure, task distribution, communication
Mode is studied in coordination.In order to enable research, with more the meaning in actual scene, researchers concentrate to some multimachine devices
People's task is studied, including is formed into columns and cooperated, search for, surrounding and seize.Wherein, it is one very typical that multi-robot Cooperation, which is surrounded and seize,
Problem, it is related to many aspects in multirobot field, including multirobot architecture, communication mode, Cooperation controlling,
The technologies such as task distribution have relatively broad application after expansion in military and industrial circle.
Document [1] has inquired into the cooperation based on potential grid method and has surrounded and seize strategy, and introducing " virtual scope " reduces road
Diameter plans number.Document [2] simulates the interaction between the antibody of B cell in immune system, it is proposed that robot behavior decision
Surround and seize algorithm so that chase robot and form effective ring of encirclement, realize that multirobot is surrounded and seize.Document [3] is proposed based on multimachine
Mode is surrounded and seize in device people study, achieves certain effect.Document [4] analyzes the critical item for successfully surrounding and seize target robot
Part, and scheme is surrounded and seize in the ambuscade for devising multirobot, achieves good effect.Document [5] proposes a kind of based on dynamic
The multirobot of prediction target trajectory tail portion point surrounds and seize algorithm, is predicted robot and realized to enclose by fitting of a polynomial
It catches.For document [6] by L-type, the design of R types, M pattern fuzzy controllers realizes that multirobot is surrounded and seize, feasible by experimental verification
Property.Document [7] proposes formation, outflanks, captures strategy, and bonding state conversion ensure that task is realized.Document [8-10] uses
It attempts simulation and utilizes Immune System and animal predation process, inquire into and realize multi-robotic task.
Bibliography:
[1]SUGAR T G,KUMAR V.Control of cooperating mobile manipulators[J]
.IEEE Transactions on Robotics and Automation.2002.18(1):94-103;
[2] multirobot On The Pursuit For Linear [J] the Wuhan University Journals of Tan Yongli, Fang Yan army based on artificial immune system,
2014,47(1):105-109。
[3]Liu Jie,Liu Shuhua,Wu Hongyan,et al.A pursuit evasion algorithm based
on hierarchical reinforcement learning[C]//International Conference on Measuring
Technology and Mechatronics Automation.Zhangjiajie:IEEE,2009:482-486;
[4] it pays bravely, a kind of multirobots of Wang Hao outstanding person surround and seize strategy [J] Central China University of Science and Technology journal (natural science edition),
2008,36 (2):26-29;
[5] Hu Jun, Zhu Qing protect and surround and seize algorithm [J] electronics based on dynamic prediction track and the multirobot surrounded and seize a little
Report, 2011,39 (11):2480-2485;
[6] after happy Cao Zhi, equal multiple autonomous robots of the based on fuzzy control coordination strategy surround and seize the Central China [J] to Yuan's good jade coke by force
University of Science and Technology's journal (natural science edition), 2011,39 (2):328-331;
[7]CAO Zhi-qiang,Zhang Bin,Wang Shuo,et al.Cooperative hunting of
multiple mobile robots in an unknown environment[J].Acta Automatica Sinica,
2003,29(4):536-543;
[8]Khan,Muhammad T,De Silva,Clarence W.Autonomous and robust multi-
robot cooperation using an artificial immune system[J].International Journal
of Robotics and Automation,2012,27(1):60-75;
[9]Gao Yunyuan,Wei Wei.Multi-robot autonomous cooperation intergrated
with immune based dynamic task allocation[C]//Proceedings of the sixth
international conference on intelligent systems design and applications,
ISDA2006,2:586-591;
[10]C.Muro,R.Escobedo,L.Spector,R.P.Coppinger.Wolf-pack(Canis lupus)
hunting strategies emerge from simple rules in computational simulations[J]
.Behavioural Processes.2011(3)。
But the above research work does not fully take into account multirobot and surrounds and seize following relevant issues in strategy:First, by
In excessively simplifying multirobot problem the actual effectiveness of strategy cannot be guaranteed.Secondly, robot is right in simulated environment
The motion mode limitation of mobile robot is excessive, and does not consider the kinematical constraint of robot, does not embody mobile robot and exists
Actual motion state in environment, it is difficult to the feasibility of verification algorithm in a practical situation.Furthermore it does not provide multirobot to enclose
The entirety caught surrounds and seize flow, can not cover multirobot and surround and seize various complex situations in task.Finally, runaway is due to perception
And the deficiency on motion mode makes runaway is relatively easy to realize to surround and seize.
Invention content:
Goal of the invention:
The present invention provide it is a kind of based on fuzzy inference system multirobot angle control surround and seize method, the purpose is to solve
It is the problems of previous.
Technical solution:
Method is surrounded and seize in multirobot angle control based on fuzzy inference system, it is characterised in that:This method uses two layers
Fuzzy reasoning realizes that multirobot surrounds and seize strategy;First layer fuzzy inference system is decision-making level, for surrounding and seize task status
It is identified, selects corresponding multirobot strategy;When the person of surrounding and seize is in search condition, the output of decision-making level's fuzzy reasoning
Search, the person of surrounding and seize execute search strategy;When the person of surrounding and seize is in proximity state, decision-making level exports Approach, i.e., close to plan
Slightly;When the person of surrounding and seize, which is in, surrounds and seize state, decision-making level exports Surround, and the person of surrounding and seize carries out surrounding and seize strategy.
The design of decision-making level's fuzzy rule is with 3 persons of surrounding and seize and target distance LieAs input;Pass through membership function
It is blurred:Membership function converts accurate input value to corresponding fuzzy set and corresponding degree of membership;LieMould
Paste collection is { S, M, the L } on section;The rule in rule base is matched again;I-th rule RiIt is as follows:
Wherein LieIndicate distance of No. i-th robot apart from runaway e,For the fuzzy of n-th of the i-th rule input
Collection;SiFor the output of rule, the strategy of Tactic selection is indicated;
According to the former piece of every rule, in corresponding fuzzy set 0~1 degree of membership is calculated using trapezoidal membership function;3
A input variable is μ by the degree of membership that membership function is blurred1e, μ2e, μ3e, then rule is matched, obtain mould
The degree of membership of paste rule is the minimum value in input variable degree of membership;
The output of each rule is overlapped according to fuzzy rule degree of membership finally, is exported;The strategy packet of output
Search strategy is included, close to strategy, surrounds and seize strategy.
Search strategy:
When the output of decision-making level's fuzzy reasoning is Search, the person of surrounding and seize executes search strategy, and specific implementation is used and searched at random
Rope mode scans for;Each robot possesses multigroup alternative left and right wheel speed, wherein containing straight trip, left-hand rotation, right-hand rotation behavior;
Robot randomly selects one of which or so wheel speed to move every time;
Close to strategy:
When decision-making level exports Approach, that is, the person's of surrounding and seize overall execution is not surrounded and seize person and is connect to runaway close to strategy
Closely;Target is approached using fuzzy inference system realization;
For the fuzzy inference system, to be close to target, the current direction of the person of surrounding and seize and mesh by the future position for the person of surrounding and seize
Drift angle α between mark is exported as input as one group or so wheel speed value;
Input is blurred by membership function, the fuzzy set of angle on -180 ° to 180 ° NM, NS, O, PS,
PM};It after blurring, is made inferences according to rule base, obtains final output;Rule is as follows:
If α is NM then Output is LRT
When drift angle α is larger (NM), output LRT, i.e., one group or so the wheel speed that robot significantly turns to the right, and by
Gradually towards target;With the reduction of drift angle, robot will be by keeping straight on rapidly close to target;
The output of every rule in fuzzy rule base is one group or so wheel speed;In the output for obtaining fuzzy inference system
One group or so wheel speed after, need to calculate the coordinate after the unit interval according to the current movable information of robot;In actual rings
In border, the not simple particle of the motion mode of robot meets the motion mode of kinematical constraint, the movement of robot
There is no mutation;Therefore, the calculating of the coordinate needs the kinematical constraint for meeting wheeled robot;
In order to preferably embody the movement of robot in simulated program, the movement of all robots is based on a following 5 groups left sides
Right wheel speed is formed;
In order to ensure that multirobot surrounds and seize the feasibility of algorithm, need to consider robot kinematics' constraint, it in this way can be true
Real simulating actual conditions preferably verify multirobot and surround and seize algorithm;The linear velocity and angular speed of wheeled robot are by left and right wheels
Fast vr, vlIt is calculated
According to linear velocity and angular speed, move distance of the t inner machine people on X, Y-direction between unit can be calculated
Dx=R (1-cos ω t) (3)
Dy=Rsin ω t (4)
Wherein R is the polar diameter of robot motion's rotation, is calculated by robot linear velocity and angular speed
D in formula (5) is the wheel diameter of wheeled robot.
xn=xc+dycosωt+dxsinωt (6)
yn=yc-dysinωt+dxcosωt (7)
θn=θc+ωt (8)
It finally can be according to the current pose (x of robotc,yc,θc) the robot position after run unit time t is calculated
Appearance (xn,yn,θn), wherein x, y, θ indicate the current abscissa of robot, ordinate and respectively currently towards angle, subscript c tables
Show current time, at the time of n is indicated after unit interval t.
Multirobot based on angle control surrounds and seize strategy:
By controlling angle between the person of surrounding and seize, preferentially strategy is surrounded and seize to what angle line nearly formed the ring of encirclement;First according to head
The movable information of secondary collected runaway carries out dynamic prediction to runaway, then using runaway's predicted motion direction as standard,
Based on the person's of surrounding and seize quantity, a plurality of angle line is uniformly established;Angle line establish after just with runaway's binding positions, no longer with runaway
The direction of motion is related, to avoid the continuous rotation of runaway from leading to the significantly swing of angle line, while also reducing multimachine
Device people continually redistributes angle line and leads to not to form the possibility effectively surrounded;
Then, the person of surrounding and seize preferentially approaches nearest angle line, the formation and holding for ensureing to surround and seize situation with this;With
The person of surrounding and seize successively after equal angle of arrival line, situation is surrounded and gradually forms;
Finally, while keeping on angle line, adjustment pose constantly close to runaway, is reduced and is surrounded each person of surrounding and seize
Circle, realization are finally surrounded and seize.
The person of surrounding and seize current location is to the line of runaway's predicted position, the folder with the angle line of runaway's predicted position foundation
When the absolute value of angle θ is larger, illustrates that the person of surrounding and seize is not formed and surround and seize situation, then by θ is controlled realize preferentially to angle line into
Row is close;When θ level off to 0 when, illustrate that the person of surrounding and seize arrived angle line, surround and seize situation and generally form, by adjusting the person's of surrounding and seize
The direction of motion will keep θ to level off to 0 while close, to ensure to surround and seize the holding of situation, to realize the receipts of the ring of encirclement
Contracting.
Prediction to runaway is to predict the position after certain step number according to the current posture information of runaway;Step number
Calculating is obtained by formula:
N=μm of ax (L1e,L2e,L3e) (9)
Wherein LieAt a distance from indicating 3 persons of surrounding and seize between runaway;When the person's of surrounding and seize overall distance target farther out when, predict step number
Larger, μ is obtained according to actual environment experiment;It is approached as robot is whole, the ring of encirclement is gradually reduced, and prediction step number gradually drops
Down to runaway's physical location.
Fuzzy inference system realizes that strategy is surrounded and seize in angle control:
Since the strategy of each robot is consistent, i.e., preferential angle of arrival line is carried out again keeping close, therefore every
A robot is all completed by a set of fuzzy inference system, can be reduced in this way since multi-robot system input quantity excessively leads to mould
The rule base for pasting inference system is excessively huge;The input of fuzzy inference system is:The angle of runaway and the person's of surrounding and seize direction of motion
It is (empty to establish angle line according to the prediction pose of runaway to the angle β of the line and angle line of runaway by α, and the person of surrounding and seize
Line);The angle of the person of surrounding and seize position and angle line is α, and the formation of situation, the person of surrounding and seize movement side are surrounded and seize by adjusting the control of α angles
Be β to the angle with runaway's direction of motion, show in a dynamic environment to the person of surrounding and seize and runaway's movable information
Consider;
Using α and β as the input of fuzzy inference system;Input variable is blurred by trapezoidal membership function;α angles
Fuzzy set serve as reasons { N, ZE, P } on -30 ° to+30 °, the fuzzy sets of β angles is by { S, the M.L.T } on 0 ° to 180 °;Root
According to the i-th rule RiIt is shown, when α is fuzzy setβ isWhen, the left and right wheel speed of rule one group of straight trip of input;
When α is larger, according to the movable information that currently do not surround and seize between person and runaway, its direction is adjusted into row major, is made
It approaches angle line;Near the person's of surrounding and seize angle of arrival line, then the movement of robot is adjusted, it is made to keep this opposite
The person of surrounding and seize is approached in the case of angle position.
For entirely surrounding and seize successful judgement, the distance by the person of surrounding and seize apart from runaway, between runaway distance and
The cluster center position that the runaway distance person of surrounding and seize is formed judges.
Advantage and effect:
The present invention provide it is a kind of based on fuzzy inference system multirobot angle control surround and seize method.There is shown herein more
The whole implementation process that robot is surrounded and seize, and task creation hierarchical mode is surrounded and seize to multirobot by fuzzy inference system.
The movement for wherein surrounding and seize robot and robot of escaping meets body movement constraint, and is that runaway devises preferably simultaneously
Policy of scuttle.Finally, pass through the feasibility of evaluated algorithm.It is in the kinematical constraint for fully considering wheeled robot
Under the premise of, it surrounds and seize robot and its movement tendency is predicted according to runaway's location information, preferentially formed by angle control
Situation is surrounded and seize, realizes that multirobot surrounds and seize control using fuzzy inference system, realizes that the formation of the ring of encirclement and final shrink are surrounded.
Finally by the validity of experimental verification research method, the results showed that the algorithm can effectively realize that multirobot is surrounded and seize.
Description of the drawings
Fig. 1 is initially to surround and seize environment map;
Fig. 2 surrounds and seize for successfully to scheme;
Fig. 3 is that surrounds and seize task overall flow figure;
Fig. 4 is to surround and seize administrative division map;
Fig. 5 is apart from membership function figure;
Fig. 6 is robot to the close figure of target point;
Fig. 7 is wheeled robot motor behavior figure;
Fig. 8 is that tactful schematic diagram is surrounded and seize in the control of angles;
Fig. 9 is fuzzy inference system input figure;
Figure 10 is the Artificial Potential Field figure based on virtual target point;
Figure 11 is that surrounds and seize successfully process decision chart;
Figure 12 is initial environment figure;
Figure 13 is random search figure;
Figure 14 is close to policy map;
Figure 15 is to surround and seize policy map;
Figure 16 is finally to surround and seize figure;
For, other original states surround and seize tactful effect to Figure 17, and wherein a is that experiments one are initial;B is that experiment one is finally surrounded and seize;
C is two initial graphs of experiment;D is that experiment two finally surrounds and seize figure;
Figure 18 is that runaway schemes with geometric center distance;
Figure 19 distance maps between the person of surrounding and seize;
Figure 20 is the person of surrounding and seize apart from runaway's distance map.
Specific implementation mode:
In conjunction with attached drawing, the present invention is described further first:
The present invention provide it is a kind of based on fuzzy inference system multirobot angle control surround and seize method, multirobot is surrounded and seize
Task description:
It is surrounded and seize for specific multirobot and recognizes task, provided and be described below:In a continuous limited region, exist more
A robot of escaping (circle) (as shown in Figure 1) for surrounding and seize robot (rectangular) and one, their random distributions are in the environment.Machine
Darker regions around device people are the sensing range of each robot of simulation.It being capable of shared information between the person of surrounding and seize.In order to protect
The maximum movement speed of card justice, runaway and the person of surrounding and seize are identical with sensing range.Therefore when the person of surrounding and seize has found runaway
While, runaway has found the person of surrounding and seize, and the single person of surrounding and seize cannot achieve and be arrested to target.
When in the sensing range that runaway appears in the arbitrary person of surrounding and seize, then it is assumed that find that runaway, finder will notify
Other teammates have found that target, multirobot start cooperation and surrounded and seize, and runaway proceeds by policy of scuttle.Wherein, it finds
Person carries out the position that target can move into next step pre- by the movable information (position, speed, direction etc.) of record target
It surveys, and other robot is notified to surround and seize the estimation range.
Ensure not collide between avoiding obstacles and robot in entire motion process.As shown in Fig. 2, examining
Distance and the person's of surrounding and seize distance are considered between runaway and the person of surrounding and seize from after the above distance reduces to a certain extent, escaping
Person thinks to surround and seize at this time since surrounding for the person of surrounding and seize leads to not mobile or when can not all be detached from the ring of encirclement to all directions movement
Success.
Multirobot surrounds and seize process:
In order to consider that multirobot surrounds and seize the complex situations being likely encountered in task comprehensively, it is proposed here complete multimachine
Device people surrounds and seize flow, as shown in Figure 3.
(1) initial environment:The multiple person's of surrounding and seize dispersions in the environment, search single runaway by random searching strategy
Rope.
(2) person of surrounding and seize is found:After having the person of surrounding and seize to find runaway, finder is according to target and acquires target movement letter
Breath, and broadcast and find that this information of target, the person of surrounding and seize integrally are proceeded by and surrounded and seize.
(3) it approaches:When the person's of surrounding and seize overall distance runaway farther out when, the person of surrounding and seize preferentially carries out close to strategy runaway,
Lead to surround and seize overlong time to avoid too early surround and seize, efficiency reduces.
(4) it surrounds and seize:When the person's of surrounding and seize overall distance target is closer, the person of surrounding and seize encloses runaway according to strategy is surrounded and seize
It catches, until surrounding and seize success.
(5) target is lost:During entire surround and seize, due to the complexity of environment and task, cause the person of surrounding and seize may
Lose runaway's target.When losing target, according to last time find target position, to target region that may be present into
Row search
Multirobot surrounds and seize strategy:
Realize that multirobot surrounds and seize strategy herein by fuzzy inference system.Fuzzy inference system is one rule-based
Intelligent controller, be mainly used to solve the problems, such as the complicated inference with blooming.Since it is desirable to robots to have people
The intelligence of class, and fuzzy inference system conforms exactly to the thinking habit of the mankind, need not establish accurate mathematical model, passes through
Linguistic variable replaces mathematical variable, is easy to realize the thinking strategy of the mankind in control aspect.Therefore fuzzy inference system is applicable in
It is controlled in complex environment, and is widely used in moveable robot movement control.
In order to ensure entirely to surround and seize efficiently accomplishing for task, need to ensure that the person of surrounding and seize is in due course to surround and seize into behavior,
Rather than simply carry out encirclement contraction.Premature surrounded and seize will lead to surround and seize success rate reduction.
As shown in figure 4, central point is runaway.Centered on runaway, it is respectively L to establish radius1And L2Sensing region
With surround and seize region.The entire task of surrounding and seize is divided into three kinds of states.
(1) search condition:When all persons of surrounding and seize are when other than sensing region, the person of surrounding and seize does not have found runaway, the person of surrounding and seize
Generally in the search condition to runaway.
(2) proximity state:Sensing region is arrived at when there is the person of surrounding and seize, then it is assumed that runaway has found each other, at this time with the person of surrounding and seize
Pay close attention to other person of surrounding and seize regions:When the person of surrounding and seize is when surrounding and seize the position other than region, illustrate the person's of surrounding and seize overall distance mesh
Farther out, the person of surrounding and seize should approach runaway mark.
(3) state is surrounded and seize:With the execution close to strategy, surrounds and seize in region or feel when there is the person of surrounding and seize more than half to be in
When knowing in region, the person of surrounding and seize will integrally surround and seize target.
Decision-making level's fuzzy inference system:
In order to avoid because the increase of robot quantity causes the regular quantity in fuzzy rule base excessive, using herein
Two layers of fuzzy reasoning realizes that multirobot surrounds and seize strategy.First layer fuzzy inference system is decision-making level, for above
It surrounds and seize task status to be identified, selects corresponding multirobot strategy.
When the person of surrounding and seize is in search condition, the output Search of decision-making level's fuzzy reasoning, the person of surrounding and seize executes search strategy;
When the person of surrounding and seize is in proximity state, decision-making level exports Approach, i.e., close to strategy;When the person of surrounding and seize, which is in, surrounds and seize state,
Decision-making level exports Surround, and the person of surrounding and seize carries out surrounding and seize strategy.
The design of decision-making level's fuzzy rule is with 3 persons of surrounding and seize and target distance LieAs input.By being subordinate to shown in Fig. 5
Degree function is blurred:
Membership function converts accurate input value to corresponding fuzzy set and corresponding degree of membership.LieFuzzy set
For { S, M, the L } on section.The rule in rule base is matched again.I-th rule RiIt is as follows:
Wherein LieIndicate distance of No. i-th robot apart from runaway e,For the fuzzy of n-th of the i-th rule input
Collection.SiFor the output of rule, the strategy of Tactic selection is indicated.
According to the former piece of every rule, in corresponding fuzzy set 0~1 degree of membership is calculated using trapezoidal membership function.3
A input variable is μ by the degree of membership that membership function is blurred1e, μ2e, μ3e, then rule is matched, obtain mould
The degree of membership of paste rule is the minimum value in input variable degree of membership.
The output of each rule is overlapped according to fuzzy rule degree of membership finally, is exported.The strategy packet of output
Search strategy is included, close to strategy, surrounds and seize strategy.
Strategic layer fuzzy inference system:
Search strategy:
When the output of decision-making level's fuzzy reasoning is Search, the person of surrounding and seize executes search strategy, and specific implementation is used and searched at random
Rope mode scans for.Each robot possesses multigroup alternative left and right wheel speed, wherein containing the row such as straight trip, left-hand rotation, right-hand rotation
For.Robot randomly selects one of which or so wheel speed to move every time.
Close to strategy:
When decision-making level exports Approach, that is, the person's of surrounding and seize overall execution is not surrounded and seize person and is connect to runaway close to strategy
Closely.Target is approached using fuzzy inference system realization.
For the fuzzy inference system, to be close to target, the current direction of the person of surrounding and seize and mesh by the future position for the person of surrounding and seize
Drift angle α between mark is exported as input as one group or so wheel speed value.As shown in fig. 6, following circumference point is target point, machine
Input of the people towards the angle with target point as fuzzy inference system.
Input is blurred by membership function, the fuzzy set of angle on -180 ° to 180 ° NM, NS, O, PS,
PM}.It after blurring, is made inferences according to rule base, obtains final output.Example rule is as follows:
Ifαis NM then Output is LRT
When drift angle α is larger (NM), output LRT, i.e., one group or so the wheel speed that robot significantly turns to the right, and by
Gradually towards target.With the reduction of drift angle, robot will be by keeping straight on rapidly close to target.
The output of every rule in fuzzy rule base is one group or so wheel speed.In the output for obtaining fuzzy inference system
One group or so wheel speed after, need to calculate the coordinate after the unit interval according to the current movable information of robot.In actual rings
In border, the not simple particle of the motion mode of robot meets the motion mode of kinematical constraint, the movement of robot
There is no mutation.Therefore, the calculating of the coordinate needs the kinematical constraint for meeting wheeled robot.
In order to preferably embody the movement of robot in simulated program, the movement of all robots is based on a following 5 groups left sides
Right wheel speed is formed.
As shown in fig. 7, wherein each arrow represents one group or so wheel speed, robot is embodied by the difference of left and right wheel speed
Steering direction and amplitude.The left and right wheel speed of the output of fuzzy inference system is all based on the linear superposition of above 5 groups of wheel speeds.
In order to ensure that multirobot surrounds and seize the feasibility of algorithm, need to consider robot kinematics' constraint, it in this way can be true
Real simulating actual conditions preferably verify multirobot and surround and seize algorithm.The linear velocity and angular speed of wheeled robot are by left and right wheels
Fast vr, vlIt is calculated
According to linear velocity and angular speed, move distance of the t inner machine people on X, Y-direction between unit can be calculated
Dx=R (1-cos ω t) (3)
Dy=Rsin ω t (4)
Wherein R is the polar diameter of robot motion's rotation, is calculated by robot linear velocity and angular speed
D in formula (5) is the wheel diameter of wheeled robot.
xn=xc+dycosωt+dxsinωt (6)
yn=yc-dysinωt+dxcosωt (7)
θn=θc+ωt (8)
It finally can be according to the current pose (x of robotc,yc,θc) the robot position after run unit time t is calculated
Appearance (xn,yn,θn)。
Multirobot based on angle control surrounds and seize strategy:
In view of the person of surrounding and seize is consistent with the maximum speed of runaway, in order to avoid excessively leading to all enclose close to runaway
The person of catching can not assume a ring of encirclement in the side of runaway, it is proposed that by angle between the control person of surrounding and seize, preferentially to angle line
Nearly form the ring of encirclement surrounds and seize strategy.First according to the movable information of collected runaway for the first time, to runaway into action
State is predicted, then using runaway's predicted motion direction as standard, based on the person's of surrounding and seize quantity, uniformly establishes a plurality of angle line.Angle line
It is no longer related to runaway's direction of motion just with runaway's binding positions after foundation, to avoid the continuous rotation of runaway from leading
The significantly swing of angle line is caused, while also reducing multirobot and angle line is continually redistributed and leads to not shape
At the possibility effectively surrounded.Then, the person of surrounding and seize preferentially approaches nearest angle line, the formation for ensureing to surround and seize situation with this
With holding.After the person of surrounding and seize successively equal angle of arrival line, surrounds situation and gradually form.Finally, each person of surrounding and seize is keeping
While on angle line, adjustment pose constantly close to runaway, reduces the ring of encirclement, and realization is finally surrounded and seize.
Fig. 8 is that three persons of surrounding and seize carry out angle line in close schematic diagram.The wherein predicted position of central point runaway, θ
For the person of surrounding and seize current location to the angle of runaway's predicted position line and immediate angle line.When the absolute value of θ is larger,
Illustrate that the person of surrounding and seize is not formed and surround and seize situation, then preferentially angle line is approached to realize by controlling θ;When θ levels off to 0
When, illustrate that the person of surrounding and seize arrived angle line, surround and seize situation and generally form, by adjusting the direction of motion for the person of surrounding and seize, close
θ is kept to level off to 0 simultaneously, to ensure to surround and seize the holding of situation, to realize the contraction of the ring of encirclement.
Prediction to runaway is to predict the position after certain step number according to the current posture information of runaway.Step number
Calculating is obtained by formula:
N=μm of ax (L1e,L2e,L3e) (9)
Wherein LieAt a distance from indicating 3 persons of surrounding and seize between runaway.When the person's of surrounding and seize overall distance target farther out when, predict step number
Larger, μ is obtained according to actual environment experiment.It is approached as robot is whole, the ring of encirclement is gradually reduced, and prediction step number gradually drops
Down to runaway's physical location.
Fuzzy inference system realizes that strategy is surrounded and seize in angle control:
Since the strategy of each robot is consistent, i.e., preferential angle of arrival line is carried out again keeping close, therefore every
A robot is all completed by a set of fuzzy inference system, can be reduced in this way since multi-robot system input quantity excessively leads to mould
The rule base for pasting inference system is excessively huge.The input of fuzzy inference system is as shown in Figure 9.According to the prediction pose of runaway,
Establish angle line (dotted line).α is the angle of the person of surrounding and seize position and angle line, and the formation of situation is surrounded and seize by adjusting the control of α angles,
β is the angle of the person's of surrounding and seize direction of motion and runaway's direction of motion, show in a dynamic environment to the person of surrounding and seize and escape
The considerations of person's movable information.
Using α and β as the input of fuzzy inference system.Input variable is blurred by trapezoidal membership function.α angles
Fuzzy set serve as reasons { N, ZE, P } on -30 ° to+30 °, the fuzzy sets of β angles is by { S, the M.L.T } on 0 ° to 180 °, mould
Paste rule is formulated according to actual experiment.Such as the i-th rule RiIt is shown, when α is fuzzy setβ isWhen, rule input one
The left and right wheel speed of group straight trip.
When α is larger, according to the movable information that currently do not surround and seize between person and runaway, its direction is adjusted into row major, is made
It approaches angle line;Near the person's of surrounding and seize angle of arrival line, then the movement of robot is adjusted, it is made to keep this opposite
The person of surrounding and seize is approached in the case of angle position.
According to the fuzzy inference system, it is ensured that the person of surrounding and seize is in the case where considering runaway's motion state, uniformly
It is distributed in around runaway.During the ring of encirclement is formed, by a small margin close to runaway;After ring of encirclement formation,
Significantly carry out contraction encirclement.
Runaway's strategy:
In order to ensure to surround and seize the validity of algorithm, runaway is also required to certain intelligent strategy and surrounds and seize strategy to cope with.
Here intelligent strategy of the Artificial Potential Field as runaway is used.After runaway detects the person of surrounding and seize, the person of surrounding and seize is considered as obstacle
Object, separate by the repulsion progress of barrier, the resultant direction that direction is repulsion of escaping, speed is the maximum speed of runaway.
As shown in Figure 10, the runaway at center is surrounded and seize by the synthesis of 2 persons of surrounding and seize.Pass through the repulsion meter of Artificial Potential Field
It calculates, the direction of escaping of runaway is right-hand arrow direction.Since there is no the fortune for considering wheeled robot for traditional artificial potential field
It is dynamic to learn constraint, so introducing the motion control that fuzzy inference system realizes runaway herein.In the maximum step on direction of escaping
A virtual target point is established in length, is to be approached to virtual target point by the conversion of motion of runaway, then by previously mentioned
It is executed close to strategy based on fuzzy inference system.
Surround and seize success conditions:
For entirely surrounding and seize successful judgement, the distance by the person of surrounding and seize apart from runaway, between runaway distance and
The cluster center position that the runaway distance person of surrounding and seize is formed judges.
In Figure 11, open circles indicate that the person of surrounding and seize, center dark color great circle indicate that runaway, roundlet indicate that the person of surrounding and seize is formed several
What center.The distance between i-th of person of surrounding and seize and j-th of person of surrounding and seize are denoted as Sij, the person of surrounding and seize formed geometric center with escape
The distance of runner is denoted as C.When the above distance is respectively less than certain threshold value (εs,ε1,εc) when, it is believed that success is surrounded and seize, specific threshold value is set
It is fixed to be set by the robot speed of service and actual environment.It can ensure the encirclement for the person of surrounding and seize by the judgement of this 3 groups of distances
The validity of circle.
In above-mentioned formula, xi,yiAnd indicate the coordinate of No. i-th robot, xe,yeIndicate runaway's coordinate.
Experimental result and analysis:
This algorithm is simulated actual multirobot by emulating and is surrounded and seize.In order to make emulation have more authenticity, acquisition is real
The related data of body robot and actual environment, including robot size, wheels of robot diameter, robot highest wheel speed, most
High safety speed and effectively perceive range, reduce, are embodied in simulated program according to a certain percentage.Surround and seize success threshold according to
Actual conditions carry out setting
Figure 12 is that multirobot is surrounded and seize the incipient stage, and 3 red persons of surrounding and seize chase the runaway of blue.Each other not
It was found that other side.Runaway carries out policy selection according to the decision-making level of Section 4, executes random searching strategy.
Such as Figure 13, according to search strategy, the person of surrounding and seize carries out random search in simulated environment.
As shown in figure 14, when the sensing range for having the person of surrounding and seize to enter intermediate runaway, and other robot is in and surrounds and seize model
When enclosing outer, the close strategy of decision-making level's selection.Close to during, the person of surrounding and seize is acquired and predicts to runaway's movable information,
White square in figure in front of runaway is predicted position, and the person of surrounding and seize is integrally close to predicted position.Due to the fortune of close strategy
Row, ensure that the direction of motion of the person's of surrounding and seize entirety towards runaway, be conducive to the progress for next surrounding and seize strategy.
After close to strategy operation a period of time, the person of surrounding and seize overall distance runaway is closer, and strategy is surrounded and seize in decision-making level's selection.
Such as Figure 15, the person of surrounding and seize integrally establishes angle line to predicted position and surrounds.Respectively angle line is close to its by each person of surrounding and seize.
By the operation of a period of time, the person of surrounding and seize finally realizes and is surrounded and seize to runaway, such as Figure 16.With constantly diagonal
The close of line is spent, situation is surrounded and seize and generally forms, person's contraction ring of encirclement completion is not finally surrounded and seize and finally surrounds and seize, runaway is equal by the person of surrounding and seize
Even being enclosed in geometric center can not move, and surround and seize success.In the whole process, it can be seen that the movement rail of arbitrary robot
Mark is consecutive variations.Robot adjusts movement by left and right wheels differential, meets the kinematical constraint of wheeled robot.
Strategy is surrounded and seize for multirobot, for the efficiency and applicability of checking research method, we are in different rings
Border carries out surrounding and seize experiment.Figure 17 is under the primary condition that the person of surrounding and seize and runaway have found each other, and originally surrounding and seize algorithm can be effective
Realization final surround and seize.
In order to analyze the validity for the process of surrounding and seize, need to consider many factors.In this experiment, it is primarily upon and surrounds and seize state
The holding of gesture and the effect for surrounding contraction.Such as Figure 11, effect is mainly surrounded and seize by the parameter evaluation of 3 aspects:LIe,Sij,C.It is comprehensive
Consider that above-mentioned tripartite's face data variation surrounds and seize strategy to evaluate during entirely surrounding and seize.
Figure 18, different lines indicates that difference surrounds and seize experiment, the situation of change of relevant evaluation parameter three times respectively in 19,20.
Abscissa is movement step number, and ordinate is the range information after normalization.It can be seen that during entire surround and seize, due to first
Beginning position is different, and the initial value of several evaluation parameters simultaneously differs, but between the person of surrounding and seize between distance and the person of surrounding and seize and runaway
Distance stablize with the increase of operation step number and reduce, embody the maintenance and contraction that situation is surrounded and seize during entire surround and seize.
Runaway at a distance from the person's of surrounding and seize geometric center there are fluctuation within a narrow range, main reason is that perceive the person of surrounding and seize laggard by runaway
Gone it is uncertain escape, but with the progress surrounded and seize, will finally arrive at a lower numerical value.
In summary:
Multirobot is had studied herein and surrounds and seize task, and task bulk flow is surrounded and seize by detailed in the analysis design to task
Journey, and task is surrounded and seize by two layers of fuzzy inference system completion.First layer carries out decision by component environment data, and selection is searched
Rope approaches or surrounds and seize strategy.The second layer realizes the multimachine device for approaching and being controlled based on angle by fuzzy inference system respectively
People surrounds and seize strategy.It is tested using simulated program, the environment of simulated program proportionally contracts according to the parameter in actual environment
Subtract, the movement of robot meets kinematics model constraint.Multi simulation running experiment is carried out under different primary condition, demonstrates calculation
The feasibility of method achieves preferable effect.
Claims (5)
1. method is surrounded and seize in the multirobot angle control based on fuzzy inference system, it is characterised in that:This method uses two layers of mould
Reasoning is pasted to realize that multirobot surrounds and seize strategy;First layer fuzzy inference system be decision-making level, for surround and seize task status into
Row identification, selects corresponding multirobot strategy;When the person of surrounding and seize is in search condition, the output of decision-making level's fuzzy reasoning
Search, the person of surrounding and seize execute search strategy;When the person of surrounding and seize is in proximity state, decision-making level exports Approach, i.e., close to plan
Slightly;When the person of surrounding and seize, which is in, surrounds and seize state, decision-making level exports Surround, and the person of surrounding and seize carries out surrounding and seize strategy;
The design of decision-making level's fuzzy rule is with 3 persons of surrounding and seize and target distance LieAs input;It is carried out by membership function
Blurring:Membership function converts accurate input value to corresponding fuzzy set and corresponding degree of membership;LieFuzzy set
For { S, M, the L } on section;The rule in rule base is matched again;I-th rule RiIt is as follows:
Wherein LieIndicate distance of No. i-th robot apart from runaway e,For the fuzzy set of n-th of input of the i-th rule;
SiFor the output of rule, the strategy of Tactic selection is indicated;
According to the former piece of every rule, in corresponding fuzzy set 0~1 degree of membership is calculated using trapezoidal membership function;3 defeated
It is μ to enter variable by the degree of membership that membership function is blurred1e, μ2e, μ3e, then rule is matched, obtain fuzzy rule
Degree of membership then is the minimum value in input variable degree of membership;
The output of each rule is overlapped according to fuzzy rule degree of membership finally, is exported;The strategy of output includes searching
Rope strategy surrounds and seize strategy close to strategy;
Multirobot based on angle control surrounds and seize strategy:
By controlling angle between the person of surrounding and seize, preferentially strategy is surrounded and seize to what angle line nearly formed the ring of encirclement;Basis is adopted for the first time first
The movable information of the runaway collected carries out dynamic prediction to runaway, then using runaway's predicted motion direction as standard, is based on
The person's of surrounding and seize quantity uniformly establishes a plurality of angle line;It is just no longer moved with runaway with runaway's binding positions after angle line foundation
Directional correlation to avoid the continuous rotation of runaway from leading to the significantly swing of angle line, while also reducing multirobot
Angle line is continually redistributed and leads to not to form the possibility effectively surrounded;
Then, the person of surrounding and seize preferentially approaches nearest angle line, the formation and holding for ensureing to surround and seize situation with this;With enclosing
The person of catching after equal angle of arrival line, surrounds situation and gradually forms successively;
Finally, for each person of surrounding and seize while keeping on angle line, adjustment pose constantly close to runaway, reduces the ring of encirclement, real
Now finally surround and seize;
Fuzzy inference system realizes that strategy is surrounded and seize in angle control:
Since the strategy of each robot is consistent, i.e., preferential angle of arrival line is carried out again keeping close, therefore each machine
Device people is completed by a set of fuzzy inference system, can be reduced in this way since multi-robot system input quantity excessively leads to fuzzy push away
The rule base of reason system is excessively huge;The input of fuzzy inference system is:The angle α of runaway and the person's of surrounding and seize direction of motion, with
And the angle β of the person of surrounding and seize to the line and angle line of runaway establishes angle line (dotted line) according to the prediction pose of runaway;
The angle of the person of surrounding and seize position and angle line is α, and the formation of situation is surrounded and seize by adjusting the control of α angles, the person's of surrounding and seize direction of motion with
The angle of runaway's direction of motion is β, show in a dynamic environment to the person of surrounding and seize and runaway's movable information the considerations of;
Using α and β as the input of fuzzy inference system;Input variable is blurred by trapezoidal membership function;The mould of α angles
The fuzzy set of { N, ZE, the P } that paste collection is served as reasons on -30 ° to+30 °, β angles are by { S, the M.L.T } on 0 ° to 180 °;According to i-th
Rule RiIt is shown, when α is fuzzy setβ isWhen, the left and right wheel speed of rule one group of straight trip of input;
When α is larger, according to the movable information that currently do not surround and seize between person and runaway, its direction is adjusted into row major, keeps its right
Angle line is approached;Near the person's of surrounding and seize angle of arrival line, then the movement of robot is adjusted, it is made to keep relative angle position
The person of surrounding and seize is approached in the case of setting.
2. method is surrounded and seize in the multirobot angle control according to claim 1 based on fuzzy inference system, feature exists
In:Search strategy:
When the output of decision-making level's fuzzy reasoning is Search, the person of surrounding and seize executes search strategy, and specific implementation uses random search side
Formula scans for;Each robot possesses multigroup alternative left and right wheel speed, wherein containing straight trip, left-hand rotation, right-hand rotation behavior;Machine
People randomly selects one of which or so wheel speed to move every time;
Close to strategy:
When decision-making level exports Approach, that is, the person's of surrounding and seize overall execution is not surrounded and seize person and is approached to runaway close to strategy;It adopts
Target is approached with fuzzy inference system realization;
For the fuzzy inference system, with by the future position for the person of surrounding and seize be close to target, the current direction of the person of surrounding and seize and target it
Between drift angle α as input, export as one group or so wheel speed value;
Input is blurred by membership function, and the fuzzy set of angle is on -180 ° to 180 °
{NM,NS,O,PS,PM};It after blurring, is made inferences according to rule base, obtains final output;Rule is as follows:
If α is NM then Output is LRT
When drift angle α is larger (NM), LRT, i.e., one group or so the wheel speed that robot significantly turns to the right, and gradual court are exported
To target;With the reduction of drift angle, robot will be by keeping straight on rapidly close to target;
The output of every rule in fuzzy rule base is one group or so wheel speed;The one of the output for obtaining fuzzy inference system
After group or so wheel speed, need to calculate the coordinate after the unit interval according to the current movable information of robot;In the actual environment,
The not simple particle of the motion mode of robot, meets the motion mode of kinematical constraint, and the movement of robot is not deposited
It is being mutated;Therefore, the calculating of the coordinate needs the kinematical constraint for meeting wheeled robot;
In order to preferably embody the movement of robot in simulated program, the movement of all robots is based on following 5 groups of left and right wheels
Speed is formed;
In order to ensure that multirobot surrounds and seize the feasibility of algorithm, need to consider robot kinematics' constraint, it in this way can true mould
Quasi- actual conditions preferably verify multirobot and surround and seize algorithm;The linear velocity and angular speed of wheeled robot are by left and right wheel speed vr,
vlIt is calculated
According to linear velocity and angular speed, move distance of the t inner machine people on X, Y-direction between unit can be calculated
Dx=R (1-cos ω t) (3)
Dy=Rsin ω t (4)
Wherein R is the polar diameter of robot motion's rotation, is calculated by robot linear velocity and angular speed
D in formula (5) is the wheel diameter of wheeled robot;
xn=xc+dycosωt+dxsinωt (6)
yn=yc-dysinωt+dxcosωt (7)
θn=θc+ωt (8)
It finally can be according to the current pose (x of robotc,yc,θc) the robot pose (x after run unit time t is calculatedn,
yn,θn), wherein x, y, θ indicate the current abscissa of robot, ordinate and respectively currently towards angles, and subscript c indicates current
Moment, n indicate unit interval t after at the time of.
3. method is surrounded and seize in the multirobot angle control according to claim 2 based on fuzzy inference system, feature exists
In:The person of surrounding and seize current location is to the line of runaway's predicted position, the angle theta for the angle line established with runaway predicted position
When absolute value is larger, illustrates that the person of surrounding and seize is not formed and surround and seize situation, then preferentially angle line is connect to realize by controlling θ
Closely;When θ level off to 0 when, illustrate that the person of surrounding and seize arrived angle line, surround and seize situation and generally form, by adjusting the movement for the person of surrounding and seize
Direction will keep θ to level off to 0 while close, to ensure to surround and seize the holding of situation, to realize the contraction of the ring of encirclement.
4. method is surrounded and seize in the multirobot angle control according to claim 3 based on fuzzy inference system, feature exists
In:
Prediction to runaway is to predict the position after certain step number according to the current posture information of runaway;The calculating of step number
It is obtained by formula:
N=μm of ax (L1e,L2e,L3e) (9)
Wherein LieAt a distance from indicating 3 persons of surrounding and seize between runaway;When the person's of surrounding and seize overall distance target farther out when, prediction step number compared with
Greatly, μ is obtained according to actual environment experiment;It is approached as robot is whole, the ring of encirclement is gradually reduced, and prediction step number continuously decreases
To runaway's physical location.
5. method is surrounded and seize in the multirobot angle control according to claim 1 based on fuzzy inference system, feature exists
In:For entirely surrounding and seize successful judgement, the distance by the person of surrounding and seize apart from runaway, distance and runaway between runaway
The cluster center position that the distance person of surrounding and seize is formed judges.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610032491.7A CN105487544B (en) | 2016-01-18 | 2016-01-18 | Method is surrounded and seize in multirobot angle control based on fuzzy inference system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610032491.7A CN105487544B (en) | 2016-01-18 | 2016-01-18 | Method is surrounded and seize in multirobot angle control based on fuzzy inference system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105487544A CN105487544A (en) | 2016-04-13 |
CN105487544B true CN105487544B (en) | 2018-10-16 |
Family
ID=55674581
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610032491.7A Expired - Fee Related CN105487544B (en) | 2016-01-18 | 2016-01-18 | Method is surrounded and seize in multirobot angle control based on fuzzy inference system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105487544B (en) |
Families Citing this family (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP6714791B2 (en) * | 2016-07-13 | 2020-07-01 | 株式会社バンダイナムコエンターテインメント | Simulation system and program |
CN108000512B (en) * | 2016-11-02 | 2021-02-02 | 香港理工大学深圳研究院 | Mobile mechanical arm system and enclosure control method |
CN108214483B (en) * | 2016-12-15 | 2020-12-11 | 香港理工大学深圳研究院 | Task dynamic tracking method and device for multi-fixed-base redundant manipulator |
CN106647808B (en) * | 2017-01-05 | 2020-02-14 | 台州施特自动化有限公司 | AUVs searching and trapping task allocation control method based on fuzzy control algorithm |
CN106773652B (en) * | 2017-01-25 | 2021-01-19 | 北京鸿智电通科技有限公司 | PID system and automatic parameter adjusting method thereof |
CN106843220B (en) * | 2017-02-27 | 2019-10-18 | 同济大学 | A kind of multi-Agent surrounds and seize-foraging behavior control method |
CN107727100B (en) * | 2017-10-16 | 2023-04-07 | 广东智爱机器人科技有限公司 | Mobile robot navigation track planning method based on distance type fuzzy reasoning |
CN107831780B (en) * | 2017-10-24 | 2020-09-22 | 佛山科学技术学院 | Multi-robot cooperation trapping method based on simulated annealing idea |
CN108115685B (en) * | 2017-12-01 | 2020-03-13 | 中国科学院自动化研究所 | Multi-robot enclosure capturing method |
CN108153310B (en) * | 2017-12-22 | 2020-11-13 | 南开大学 | Mobile robot real-time motion planning method based on human behavior simulation |
CN109079792A (en) * | 2018-09-05 | 2018-12-25 | 顺德职业技术学院 | A kind of target based on multirobot surrounds and seize method and system |
CN109696084B (en) * | 2019-01-04 | 2020-09-15 | 北京邮电大学 | Multi-satellite game capture strategy |
CN111310581B (en) * | 2020-01-17 | 2023-04-25 | 杭州电子科技大学 | Crowd safety assessment method based on fuzzy reasoning |
CN113095500B (en) * | 2021-03-31 | 2023-04-07 | 南开大学 | Robot tracking method based on multi-agent reinforcement learning |
CN114117768A (en) * | 2021-11-22 | 2022-03-01 | 天津大学 | Multi-agent distributed enclosure method with uncertain escaper positions |
CN117093021B (en) * | 2023-10-19 | 2024-01-30 | 西北工业大学深圳研究院 | Distributed formation surrounding method applied to group intelligent system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101059700A (en) * | 2007-03-23 | 2007-10-24 | 北京大学 | Underwater bionic robot cooperated transportation method |
CN103792846A (en) * | 2014-02-18 | 2014-05-14 | 北京工业大学 | Robot obstacle avoidance guiding method based on Skinner operating condition reflection principle |
CN104942807A (en) * | 2015-04-16 | 2015-09-30 | 上海大学 | Method for capturing targets by aid of multiple robots on basis of extensive cooperative games |
CN105182973A (en) * | 2015-09-08 | 2015-12-23 | 郑州大学 | Self-adaptive hunting device using multiple robot pursuers to hunt single moving target and method |
-
2016
- 2016-01-18 CN CN201610032491.7A patent/CN105487544B/en not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101059700A (en) * | 2007-03-23 | 2007-10-24 | 北京大学 | Underwater bionic robot cooperated transportation method |
CN103792846A (en) * | 2014-02-18 | 2014-05-14 | 北京工业大学 | Robot obstacle avoidance guiding method based on Skinner operating condition reflection principle |
CN104942807A (en) * | 2015-04-16 | 2015-09-30 | 上海大学 | Method for capturing targets by aid of multiple robots on basis of extensive cooperative games |
CN105182973A (en) * | 2015-09-08 | 2015-12-23 | 郑州大学 | Self-adaptive hunting device using multiple robot pursuers to hunt single moving target and method |
Non-Patent Citations (2)
Title |
---|
基于角度优先的多机器人围捕策略;陈阳舟;《北京工业大学学报》;20120531;第38卷(第5期);第716-717页 * |
多机器人协调作业策略研究;何威;《中国优秀硕士学位论文全文数据库》;20090315(第3期);正文第15-40页 * |
Also Published As
Publication number | Publication date |
---|---|
CN105487544A (en) | 2016-04-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105487544B (en) | Method is surrounded and seize in multirobot angle control based on fuzzy inference system | |
Zhao et al. | Survey on computational-intelligence-based UAV path planning | |
CN106671982B (en) | Driverless electric automobile automatic overtaking system system and method based on multiple agent | |
CN106483958B (en) | A kind of man-machine coordination formation based on obstacle figure and potential field method follows and barrier-avoiding method | |
Duan et al. | Optimal formation reconfiguration control of multiple UCAVs using improved particle swarm optimization | |
CN107168324A (en) | A kind of robot path planning method based on ANFIS fuzzy neural networks | |
CN105182973A (en) | Self-adaptive hunting device using multiple robot pursuers to hunt single moving target and method | |
Shang et al. | Deep reinforcement learning with reference system to handle constraints for energy-efficient train control | |
CN110716562A (en) | Decision-making method for multi-lane driving of unmanned vehicle based on reinforcement learning | |
Hong et al. | A modified dynamic window approach to obstacle avoidance combined with fuzzy logic | |
Allou et al. | Fuzzy logic controller for autonomous vehicle path tracking | |
Li et al. | Dynamically integrated spatiotemporal‐based trajectory planning and control for autonomous vehicles | |
Wei | A new formation control strategy based on the virtual-leader-follower and artificial potential field | |
Zennir et al. | Comparison of PID and fuzzy controller for path tracking control of autonomous electrical vehicles | |
Pierre | End-to-end deep learning for robotic following | |
Xu et al. | An actor-critic based learning method for decision-making and planning of autonomous vehicles | |
Zhang et al. | Multi-target encirclement with collision avoidance via deep reinforcement learning using relational graphs | |
CN115993845A (en) | Coordinated motion planning and formation control method for cluster intelligent system | |
CN113759935B (en) | Intelligent group formation mobile control method based on fuzzy logic | |
Baba et al. | A fuzzy logic-based stabilization system for a flying robot, with an embedded energy harvester and a visual decision-making system | |
Chen et al. | Non-trap artificial potential field based on virtual obstacle | |
Yuan et al. | A centralised training algorithm with D3QN for scalable regular unmanned ground vehicle formation maintenance | |
Hamdani et al. | Deep Reinforcement Learning Modeling of a V2V Communication-Based Bike Avoidance Protocol for Increased Vehicular Flow | |
Feng et al. | Mact: Multi-agent collision avoidance with continuous transition reinforcement learning via mixup | |
de Castro et al. | A vehicle following controller for highly-actuated vehicles |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20181016 Termination date: 20190118 |
|
CF01 | Termination of patent right due to non-payment of annual fee |